This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM98640QML-SP: SH mode reference clamping and evaluation board wiring

Part Number: LM98640QML-SP
Other Parts Discussed in Thread: LM98640CVAL

Hello, I'm prototyping a CCD readout board (eventual custom design) using the LM98640CVAL. The reference level does come out each line (for a couple sample clocks duration) shortly after the LSYNC so we derived an external CLAMP signal to go high at that time (the level is outside the control range of the VCLP DAC so I couldn't set it manually). We get something that sorta works, though in feeding 4 video signals (two eval boards and using both converters) I see some significant differences in the A,B,C,D columns (especially between the two boards). I'm just a bit concerned that we don't have some register settings right (e.g., should the VCLP register 0x04 be a value of 0x00 or 0x40, or does that bit not matter?). Also, do we need to adjust the jumpers on the board (the VCLP pin connected to OS{1,2}+? I worry that is providing an undesired coupling between channels 1 and 2 or otherwise biasing the capture of the reference level. Do I expect to see large gain/offset differences between individual chips? Right now I have all the settings (PGA gains, etc.) common to all but am seeing differences in the "black level" of around 1000 counts. That seems huge. Any tips?

  • Hi,

    The apps engineer most knowledgeable with this device returns to the office tomorrow at which time we will attempt to answer your questions.

    Thanks

    Christian

  • Hi Martin,

    The LM98640 is a very versatile part that can support many different types of imagining chips and operates in different modes.  The LM98640CVAL board was designed to support all the possible different operating conditions and has many different jumpers and components.   It is necessary to go through the input network and place or remove jumpers at appropriate places.  This should be sufficient to get a good input signal, but you can also remove unnecessary components. 

    If you are not using the VCLP DAC, you should power it down and disconnect the jumpers to the inputs.

    For the same input signals, you should see only minor offsets between channels and parts.  Depending upon the setup, I typically see less than 10 LSB.  The most common cause I have seen for large offsets in readings is that the clamp and sample windows have not been set correctly to capture the flat part of the signal.

    There is a GUI that allows easy programming of the registers.  You can download it from the LM98640CVAL product page on ti.com.   Unfortunately it only works with Windows XP and older versions and does not currently work with Windows 7 or 10.

    To help you further, I would need to get more details on your set up and how you are using the part.  For instance, is this CDS mode or S/H mode?   Your situation is probably too complex to handle through e2e, so I have sent you a friend request and you can contact me directly.  Note that I will be out of the office Monday through Wednesday but will still try to respond during that time if you send me something.

    Kirby